CN109754396A - Method for registering, device, computer equipment and the storage medium of image - Google Patents
Method for registering, device, computer equipment and the storage medium of image Download PDFInfo
- Publication number
- CN109754396A CN109754396A CN201811637721.8A CN201811637721A CN109754396A CN 109754396 A CN109754396 A CN 109754396A CN 201811637721 A CN201811637721 A CN 201811637721A CN 109754396 A CN109754396 A CN 109754396A
- Authority
- CN
- China
- Prior art keywords
- image
- registration
- floating
- reference picture
- anatomic landmarks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
- G06T7/0014—Biomedical image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10104—Positron emission tomography [PET]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30204—Marker
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
This application involves a kind of method for registering of image, device, computer equipment and storage mediums.Obtain reference picture and floating image to be registered;The extraction that semantic information is carried out to the reference picture and the floating image obtains including the label reference picture of institute's semantic information and marking floating image;According to institute's semantic information, the label reference picture and the corresponding target image registration model of the label floating image are determined from preset image registration model;According to institute's semantic information and the target image registration model, image registration is carried out to the reference picture and the floating image.This method solve the limitations that can only be registrated in the prior art based on single semantic information to reference picture and floating image, substantially increase the scope of application of image registration.
Description
Technical field
This application involves technical field of image processing, set more particularly to the method for registering, device, computer of a kind of image
Standby and storage medium.
Background technique
Two width or several figures that will be obtained under different time, different imaging devices or different condition may be implemented in image registration
As being matched and be superimposed, such as can to CT scan (Computed Tomography, CT) image and just
The images such as electron emission type computerized tomograph (Positron Emission Computed Tomography, PET) image
It is matched and is superimposed, to show the information for the CT image for participating in registration and the information of PET image on the same image, for clinic
Medical diagnosis provides preferable booster action, is a key technology in field of image processing.
In traditional technology, if area-of-interest (Region Of Interest, ROI) is irregular area, then extract
Irregular area in image subject to registration, and region is not advised based on this and is registrated;If ROI is key point, extract wait match
Key point in quasi- image, and be registrated based on the key point.
But in traditional technology when carrying out image registration, can only based on single semantic information such as irregular area or
Key point etc. is registrated image subject to registration, causes the scope of application of traditional method for registering lower.
Summary of the invention
Based on this, it is necessary to can only be based on single semantic information such as irregular area or key point etc. for traditional technology
Image subject to registration is registrated, the problem for causing the scope of application of traditional method for registering lower provides a kind of image registration
Method, apparatus, computer equipment and storage medium.
In a first aspect, the embodiment of the present application provides a kind of method for registering images, this method may include:
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtains including institute's semantic information
Mark reference picture and label floating image;
According to institute's semantic information, the label reference picture and the label are determined from preset image registration model
The corresponding target image registration model of floating image;
According to institute's semantic information and the target image registration model, to the reference picture and the floating image into
Row image registration.
Institute's semantic information includes: the cut zone and anatomy mark of the floating image in one of the embodiments,
At least one of note point and the reference picture cut at least one of region and anatomy mark point;It is described default
Image registration model include the image registration model based on segmentation and the registration model based on anatomic landmarks point.
In one of the embodiments, when the target image registration model is the registration based on anatomic landmarks point
It is described according to institute's semantic information and the target image registration model when model, the reference picture and the floating are schemed
As carrying out image registration, comprising:
Obtain it is described label reference picture it is subject to registration with reference to anatomic landmarks point set and it is described label floating image to
It is registrated floating anatomic landmarks point set;
According to described subject to registration with reference to anatomic landmarks point set, the floating anatomic landmarks point set subject to registration and the base
In the registration model of anatomic landmarks point, image registration is carried out to the reference picture and the floating image.
In one of the embodiments, it is described according to it is described it is subject to registration with reference to anatomic landmarks point set, it is described subject to registration floating
Dynamic anatomic landmarks point set and the registration model based on anatomic landmarks point, to the reference picture and the floating image
Carry out image registration, comprising:
It is concentrated according to the reference anatomic landmarks point set subject to registration and the floating anatomic landmarks point subject to registration each
The matching result of the title of mark point determines mark point intersection;
According to the mark point intersection, from the reference anatomic landmarks point set subject to registration and the floating dissection subject to registration
It learns mark point and concentrates determining initial reference anatomic landmarks point set and initial floating anatomic landmarks point set respectively;
According to the initial reference anatomic landmarks point set, the initial floating anatomic landmarks point set and it is described based on solution
The registration model for learning mark point is cutd open, image registration is carried out to the reference picture and the floating image.
In one of the embodiments, when the target image registration model is the image registration model based on segmentation
When, then it is described according to institute's semantic information and the target image registration model, to the reference picture and the floating image
Carry out image registration, comprising:
Obtain the corresponding segmentation reference picture of label reference picture and the corresponding segmentation floating figure of the floating image
Picture;
It is right according to the segmentation reference picture, the segmentation floating image and the image registration model based on segmentation
The reference picture and the floating image carry out image registration.
In one of the embodiments, the method also includes:
It obtains and carries out the registration result after image registration to the reference picture and the floating image;
According to the registration result and preset image integration model, image integration is carried out to the registration result.
After in one of the embodiments, described to the reference picture and floating image progress image registration,
The method also includes:
Obtain the object transformation matrix;
According to the object transformation matrix, down-sampling operation is carried out to the reference picture after obtained down-sampling with reference to figure
Picture and obtained down-sampling floating image after down-sampling operation is carried out to the floating image, determines the down-sampling reference picture
Similarity measure values between the corresponding transformed floating image of the down-sampling floating image;
At least one in translation, rotation process, tilt operation and zoom operations is carried out to the object transformation matrix
The corresponding initial parameter of the object transformation matrix is extracted in a operation;
According to the similarity measure values, the initial parameter and preset gradient descent method, target component is determined.
Second aspect, the embodiment of the present application provide a kind of image registration device, the apparatus may include:
First obtains module, for obtaining reference picture and floating image to be registered;
First extraction module is obtained for carrying out the extraction of semantic information to the reference picture and the floating image
Label reference picture and label floating image including institute's semantic information;
First determining module, for determining the label from preset image registration model according to institute's semantic information
Reference picture and the corresponding target image registration model of the label floating image;
Registration module, for according to institute's semantic information and the target image registration model, to the reference picture and
The floating image carries out image registration.
The third aspect, the embodiment of the present application provide a kind of computer equipment, and the computer equipment includes memory and place
Device is managed, the memory is stored with computer program, the following steps when processor executes the computer program:
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtains including institute's semantic information
Mark reference picture and label floating image;
According to institute's semantic information, the label reference picture and the label are determined from preset image registration model
The corresponding target image registration model of floating image;
According to institute's semantic information and the target image registration model, to the reference picture and the floating image into
Row image registration.
Fourth aspect, the embodiment of the present application provide a kind of readable storage medium storing program for executing, are stored thereon with computer program, the meter
Calculation machine program realizes following steps when being executed by processor:
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtains including institute's semantic information
Mark reference picture and label floating image;
According to institute's semantic information, the label reference picture and the label are determined from preset image registration model
The corresponding target image registration model of floating image;
According to institute's semantic information and the target image registration model, to the reference picture and the floating image into
Row image registration.
In method for registering images provided in this embodiment, device, computer equipment and readable storage medium storing program for executing, computer equipment
Available reference picture to be registered and floating image;And the extraction of semantic information is carried out to reference picture and floating image,
Obtain include semantic information label reference picture and label floating image;And then according to semantic information, match from preset image
Label reference picture target image registration model corresponding with label floating image is determined in quasi-mode type;Finally according to semanteme
Information and target image registration model carry out image registration to label reference picture and label floating image.In the present embodiment, meter
The semantic information of reference picture and floating image can first be extracted by calculating machine equipment, thus according to different semantic informations, using not
Same target image registration model is registrated reference picture and floating image, to complete the reference including a variety of semantic informations
The registration of image and floating image, solving in the prior art can only be based on single semantic information to reference picture and floating figure
As the limitation being registrated, the scope of application of image registration is substantially increased.
Detailed description of the invention
Fig. 1 is the schematic diagram of internal structure for the computer equipment that one embodiment provides;
Fig. 2 is the method for registering images flow diagram that one embodiment provides;
Fig. 3 is the method for registering images flow diagram that another embodiment provides;
Fig. 4 is the method for registering images flow diagram that another embodiment provides;
Fig. 5 is the method for registering images flow diagram that another embodiment provides;
Fig. 6 is the method for registering images flow diagram that another embodiment provides;
Fig. 7 is the image registration device structural schematic diagram that one embodiment provides;
Fig. 8 is the image registration device structural schematic diagram that another embodiment provides;
Fig. 9 is the image registration device structural schematic diagram that another embodiment provides.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Method for registering images provided by the embodiments of the present application can be adapted for computer equipment as shown in Figure 1.The calculating
Machine equipment includes processor, the memory connected by system bus, is stored with computer program in the memory, processor is held
The step of following methods embodiment can be executed when the row computer program.Optionally, which can also include net
Network interface, display screen and input unit.Wherein, the processor of the computer equipment is for providing calculating and control ability.The meter
The memory for calculating machine equipment includes non-volatile memory medium, built-in storage, which is stored with operation system
System and computer program.The built-in storage provides for the operation of operating system and computer program in non-volatile memory medium
Environment.The network interface of the computer equipment is used to communicate with external terminal by network connection.Optionally, which sets
It is standby to can be server, it can be personal computer (personal computer, PC), can also be personal digital assistant, also
It can be other terminal devices, such as tablet computer (portable android device, PAD), mobile phone etc., may be used also
To be cloud or remote server, the embodiment of the present application is to the concrete form of computer equipment and without limitation.
It should be noted that method for registering images provided by the embodiments of the present application, executing subject can be image registration
Device, the image registration device can be implemented as computer equipment by way of software, hardware or software and hardware combining
It is some or all of.In following methods embodiment, it is illustrated so that executing subject is computer equipment as an example.
Fig. 2 is the method for registering images flow diagram that one embodiment provides.What is involved is computers to set for the present embodiment
It is standby to determine image registration model according to the semantic information that reference picture and floating image extract, and to reference picture and float
The realization process of motion video progress image registration.As shown in Fig. 2, this method may include:
S202 obtains reference picture and floating image to be registered.
Specifically, reference picture and floating image can be the image of same mode, it is also possible to the image of anomalous mode state, than
Such as, reference picture and floating image can be CT image, can also be CT image with one, the other is PET image.It is optional
, computer equipment can be registrated two width or multiple image of acquisition, for example will wherein piece image be used as with reference to figure
Picture, other images are mapped to reference picture as floating image, by floating image, to realize that reference picture and floating image are solving
Cut open the alignment learned under structure.Optionally, reference picture and floating image can be the image of same individual, be also possible to different
The image of body can be the anatomical structure for including image all the same, be also possible to comprising the identical anatomical structure in part
Image, the present embodiment is to the source of reference picture and floating image and without limitation.Optionally, reference picture and floating image
It can be two dimensional image, be also possible to 3-D image, the present embodiment is to this and is not specifically limited.
S204 carries out the extraction of semantic information to the reference picture and the floating image, obtains including the semanteme
The label reference picture and label floating image of information.
Specifically, can have been instructed according to preset after computer equipment gets the reference picture and floating image of input
The neural network model perfected extracts the semantic information in reference picture and floating image, for example, if detecting lung
The corresponding region in portion, computer equipment can just come out the corresponding region segmentation of lung, to extract the corresponding language of lung
Adopted information: if detecting bone, just the corresponding position mark of bone is come out with mark point, to mention the corresponding language of bone
Adopted information: anatomic landmarks point.Computer equipment carries out reference picture and floating image using preset neural network model
It is available comprising the label reference picture and label floating image of the semantic information extracted after language message is extracted.
S206 determines the label reference picture and institute according to institute's semantic information from preset image registration model
State the corresponding target image registration model of label floating image.
Specifically, above-mentioned image registration model is for the label reference picture and label obtained after extraction semantic information
The model that floating image is registrated, such as surface matching algorithm, mutual information method, orthonormalization matrix method and least square method
Etc. corresponding algorithm model.For the label reference picture comprising different semantic informations and label floating image, computer equipment
It can use different registration models to be registrated the two, i.e., the label reference picture including cut zone and label, which float, schemes
Picture and the label reference picture and label floating image for including anatomic landmarks point, can correspond to different image registration models.
Optionally, upper semantic information includes: at least one of cut zone and anatomy mark point of floating image,
And reference picture cuts at least one of region and anatomy mark point.Wherein, upper semantic information can be for reference to figure
Anatomic landmarks point in picture and floating image, or the cut zone in reference picture and floating image.Further,
Above-mentioned anatomic landmarks point can be geometric markers point, such as gray scale extreme value or linear structure intersection point, be also possible in anatomic form
It is upper high-visible and can pinpoint anatomical landmarks, such as the key signature point or characteristic point of tissue, organ or lesion;
Above-mentioned cut zone can be reference picture and the corresponding curve of floating image or curved surface etc., such as lung, liver or region of disorder
Domain.
Optionally, above-mentioned preset image registration model may include image registration model based on segmentation and based on dissection
Learn the registration model of mark point.Further, image registration model based on segmentation be can be to including above-mentioned cut zone
Mark reference picture and mark floating image carry out image registration image registration model, as surface matching algorithm, mutual information method,
The corresponding algorithm model of the methods of gray scale average variance method;Registration model based on anatomic landmarks point is can be to including above-mentioned solution
It cuts open the label reference picture for learning mark point and marks the registration model of floating image progress image registration, as singular value decomposition is calculated
The corresponding algorithm model of the methods of method, iteration closest approach method, orthonormalization matrix method.
S208, according to institute's semantic information and the target image registration model, to the reference picture and the floating
Image carries out image registration.
Specifically, computer equipment can choose corresponding target image registration model, right according to the difference of semantic information
Reference picture and floating image carry out image registration.Optionally, it can be wrapped simultaneously in a width reference picture or a width floating image
Cut zone and anatomy point are included, at this point, computer equipment can be first with the corresponding target image registration model of anatomy point
Anatomy point in reference picture and floating image is registrated, the corresponding target image registration model of cut zone is recycled
Cut zone in reference picture and floating image is registrated;It can also match first with the corresponding target image of cut zone
Quasi-mode type is registrated the cut zone in reference picture and floating image, and the corresponding target image of anatomy point is recycled to match
Quasi-mode type is registrated the anatomy point in reference picture and floating image, can also utilize the corresponding mesh of anatomy point simultaneously
Logo image registration model is registrated the anatomy point in reference picture and floating image, and utilizes the corresponding mesh of cut zone
Logo image registration model is registrated the cut zone in reference picture and floating image, and the present embodiment does not limit this
It is fixed.
Optionally, computer equipment can be at the related operation for ensuring to continue to use CPU progress image registration therein
In the case where reason, may be incorporated into support parallel computation framework (Compute Unified Device Architecture,
CUDA graphics processor (Graphics Processing Unit, GPU)) handles partial arithmetic, above-mentioned to further speed up
To the speed for the registration Algorithm that reference picture and floating image are registrated.
Method for registering images provided in this embodiment, the available reference picture to be registered of computer equipment and the figure that floats
Picture;And the extraction of semantic information is carried out to reference picture and floating image, obtain include semantic information label reference picture with
Mark floating image;And then according to semantic information, determine that label reference picture and label are floating from preset image registration model
The corresponding target image registration model of motion video;Finally according to semantic information and target image registration model, label is joined
It examines image and label floating image carries out image registration.In the present embodiment, computer equipment can first extract reference picture and float
The semantic information of motion video, thus according to different semantic informations, using different target image registration models to reference picture
It is registrated with floating image, includes the reference picture of a variety of semantic informations and the registration of floating image to complete, solve existing
There is the limitation that can only be registrated based on single semantic information to reference picture and floating image in technology, substantially increases
The scope of application of image registration.
Fig. 3 is the method for registering images flow diagram that another embodiment provides.What is involved is work as target for the present embodiment
When image registration model is the above-mentioned registration model based on anatomic landmarks point, computer equipment is according to based on anatomic landmarks point
Registration model and semantic information process that reference picture and floating image are registrated.On the basis of the above embodiments,
Optionally, above-mentioned S208 may include:
S302 obtains the subject to registration of label reference picture and refers to anatomic landmarks point set and the label floating image
Floating anatomic landmarks point set subject to registration.
Specifically, above-mentioned reference anatomic landmarks point set subject to registration and floating anatomic landmarks point set subject to registration are each solution
Cut open the set for learning the coordinate information of mark point.Optionally, anatomic landmarks point can be the artificial mark point for carrying out preliminary making.
S304, according to it is described it is subject to registration with reference to anatomic landmarks point set, the floating anatomic landmarks point set subject to registration and
The registration model based on anatomic landmarks point carries out image registration to the reference picture and the floating image.
Specifically, the above-mentioned registration model based on anatomic landmarks point can be singular value decomposition algorithm, iteration closest approach
Any one in the corresponding algorithm model of the methods of algorithm, orthonormalization matrix method.Computer equipment can be according to acquisition
It is subject to registration with reference to anatomic landmarks point set, floating anatomic landmarks point set subject to registration and preset based on anatomic landmarks point
Registration model carries out image registration to reference picture and floating image.
Optionally, above-mentioned S304 can specifically include: according to it is above-mentioned it is subject to registration with reference to anatomic landmarks point set and it is above-mentioned to
Registration floating anatomic landmarks point concentrates the matching result of the title of each mark point, determines mark point intersection;According to above-mentioned mark
Note point intersection is concentrated true respectively from above-mentioned reference anatomic landmarks point set subject to registration and above-mentioned floating anatomic landmarks point subject to registration
Determine initial reference anatomic landmarks point set and initial floating anatomic landmarks point set;According to above-mentioned initial reference anatomic landmarks point
Collection, above-mentioned initial floating anatomic landmarks point set and the above-mentioned registration model based on anatomic landmarks point, to above-mentioned reference picture
Image registration is carried out with above-mentioned floating image.
Wherein, each anatomic landmarks point has unique title, for subject to registration with reference to anatomic landmarks point set and wait match
The mark point that the identical anatomic landmarks point of the anatomic landmarks point title that quasi- floating anatomic landmarks point is concentrated constitutes the two is handed over
Collection.Optionally, computer equipment can also refer to anatomic landmarks point set and floating anatomic landmarks point subject to registration for subject to registration
The anatomic landmarks point of concentration numbers mark point intersection of the identical anatomic landmarks point as the two.Determine mark point intersection
Afterwards, computer equipment can concentrate the corresponding point set of above-mentioned mark point intersection as initial for subject to registration with reference to anatomic landmarks point
With reference to anatomic landmarks point set, and chooses floating anatomic landmarks point subject to registration and concentrate the corresponding point set of above-mentioned mark point intersection
As initial floating anatomic landmarks point set, so as to by initial reference anatomic landmarks point set and initial floating anatomy mark
Note point set inputs the preset registration model based on anatomic landmarks point, realizes reference picture and floating image in same anatomical
Alignment under structure.
In the step of above-mentioned S304, computer equipment can according to from subject to registration with reference to anatomic landmarks point set and subject to registration
Floating anatomic landmarks point concentrates the initial reference anatomic landmarks point set and initial floating anatomic landmarks point set chosen, and benefit
Image registration is carried out to the reference picture and the floating image with the registration model based on anatomic landmarks point.Optionally,
Image registration is carried out to reference to the reference picture and the floating image using the registration model based on anatomic landmarks point
The process that image and floating image carry out image registration can be divided into the registration process of three phases, and each stage is available right
The registration process of the registration result answered, three phases is as follows:
The registration process of first stage may refer to S3042 to S3046:
S3042, according to initial reference anatomic landmarks point set and initial floating anatomic landmarks point set and it is described based on solution
The registration model for learning mark point is cutd open, determines the first registration result;First registration result include the first registration result point set and
First transformation matrix.
Specifically, computer equipment inputs initial reference anatomic landmarks point set and initial floating anatomic landmarks point set
After the preset registration model based on anatomic landmarks point, available floating anatomic landmarks point set subject to registration carries out space change
The first registration result point set and the first transformation matrix after changing.Above-mentioned first registration result point set and the first transformation matrix constitute the
One registration result.
S3044 determines the first space in the preset ratio according to the first space length set and preset ratio
Apart from corresponding first floating anatomic landmarks point set;Wherein, in the first space length set record have it is described subject to registration
The first space length of each correspondence markings point is concentrated with reference to anatomic landmarks point set and the first registration result point.
Specifically, computer equipment can be according to formula D1=after obtaining the first registration result point set | | Pf1–Pre1|
|2, calculate the first space subject to registration that each correspondence markings point is concentrated with reference to anatomic landmarks point set and the first registration result point
Distance D1, wherein Pf1Mark point structure corresponding with the first registration result point concentration is concentrated with reference to anatomic landmarks point to be subject to registration
At point set, Pre1For the first registration result point set.Optionally, above-mentioned preset ratio can need to set according to (0,1]
Interior arbitrary value.Optionally, the corresponding first floating anatomy of the first space length in preset ratio can directly be chosen
Mark point set, ascending sort can also be carried out to each distance in the first space length, then choose in preset ratio the
The corresponding first floating anatomic landmarks point set of one space length is registrated with reference to anatomic landmarks point set with first due to subject to registration
Result points concentrate the first space length of each correspondence markings point smaller, and registration result precision is higher, therefore, to the first space away from
It carries out choosing after ascending sort the first space length corresponding first in preset ratio with a distance from each in again to float solution
Label point set is cutd open, the accuracy of registration can be improved.Above-mentioned first floating anatomic landmarks point set is to solve from floating subject to registration
Cut open the corresponding point set of the first space length learned in the preset ratio that mark point concentration is chosen.
S3046, when the number for the mark point that the first floating anatomic landmarks point is concentrated is less than the preset number
When threshold value, then using first transformation matrix as object transformation matrix.
Specifically, above-mentioned object transformation matrix is that label reference picture and label floating image carry out used in image registration
Matrix, computer equipment can use object transformation matrix and realize label reference picture and mark the registration of floating image.It is optional
, the number for the mark point that computer equipment can concentrate the first floating anatomic landmarks point and preset quantity threshold carry out
Compare, is determined whether according to comparison result using above-mentioned first transformation matrix as object transformation matrix.Optionally, above-mentioned preset
Quantity threshold can be 5.When the number for the mark point that above-mentioned first floating anatomic landmarks point is concentrated is less than preset number threshold
When value, then using the first transformation matrix as object transformation matrix, and S30422 is continued to execute.
When the number for the mark point that above-mentioned first floating anatomic landmarks point is concentrated is not less than preset quantity threshold, need
Carry out the registration process of second stage.
The registration process of second stage may refer to S3048 to S30416:
S3048 is obtained described subject to registration with reference to anatomic landmarks point concentration and the first floating anatomic landmarks point set
Corresponding first refers to anatomic landmarks point set.
In this step, above-mentioned first reference anatomic landmarks point set is the label subject to registration concentrated with reference to anatomic landmarks point
The title or number of point are corresponding with the title for the label that the first floating anatomic landmarks point is concentrated or the identical mark point of number
The point set that mark point is constituted.
S30410, according to described first with reference to anatomic landmarks point set, the first floating anatomic landmarks point set and institute
The registration model based on anatomic landmarks point is stated, determines the second transformation matrix.
Specifically, identical with the method for above-mentioned the first transformation matrix of determination, computer equipment can be by first with reference to dissection
It learns label point set and the first floating anatomic landmarks point set inputs the preset registration model based on anatomic landmarks point, thus
To the second transformation matrix.
S30412 determines that second matches according to second transformation matrix and the floating anatomic landmarks point set subject to registration
Quasi- result point set.
In this step, computer equipment can be according to obtained the second transformation matrix and floating anatomic landmarks point subject to registration
The product of collection carries out spatial alternation to floating anatomic landmarks point set subject to registration using the second transformation matrix, and combines interpolation method
Such as the methods of neighbour's interpolation, bilinear interpolation or Tri linear interpolation, the second registration result point set is obtained.
S30414 is determined according to second space distance set and preset distance threshold less than described preset apart from threshold
The second space of value is apart from corresponding second floating anatomic landmarks point set;Record has described in the second space distance set
The second space distance subject to registration that each correspondence markings point is concentrated with reference to anatomic landmarks point set and the second registration result point.
In this step, after obtaining the second registration result point set, computer equipment can be according to formula D2=| | Pf–Pre2|
|2, calculate the second space subject to registration that each correspondence markings point is concentrated with reference to anatomic landmarks point set and the second registration result point
Distance D2, wherein Pf2Concentrate each mark point corresponding with the second registration result point with reference to anatomic landmarks point set to be subject to registration
Point set, Pre2For the second registration result point set.Optionally, above-mentioned preset distance threshold can be set as needed, such as distance
Threshold value can be acceptable subject to registration each corresponding with reference to anatomic landmarks point set and the second registration result point concentration according to user
The actual range of mark point is determined.Above-mentioned second floating anatomic landmarks point set is from floating anatomic landmarks point subject to registration
Concentrate the second space in the preset distance threshold chosen apart from corresponding point set.
S30416, when the number for the mark point that the second floating anatomic landmarks point is concentrated is less than the preset threshold value
When number, then using second transformation matrix as the object transformation matrix.
In this step, the number for the mark point that computer equipment can concentrate the second floating anatomic landmarks point and default
Quantity threshold be compared, determined whether according to comparison result using above-mentioned second transformation matrix as object transformation matrix.When
When the number for the mark point that above-mentioned second floating anatomic landmarks point is concentrated is less than preset quantity threshold, then square is converted by second
Battle array is used as object transformation matrix, and continues to execute S30422.
When the number for the mark point that above-mentioned second floating anatomic landmarks point is concentrated is not less than preset quantity threshold, need
Carry out the registration process of phase III.
The registration process of phase III may refer to S30418 to S30420:
S30418 is obtained described subject to registration with reference to anatomic landmarks point concentration and the second floating anatomic landmarks point set
Corresponding second refers to anatomic landmarks point set.
In this step, second with reference to anatomic landmarks point set be from it is subject to registration with reference to anatomic landmarks point concentrate choose with
The title of mark point or the corresponding point set of the identical mark point of number in above-mentioned second floating anatomic landmarks point set.
S30420, according to described second with reference to anatomic landmarks point set, the second floating anatomic landmarks point set and institute
The registration model based on anatomic landmarks point is stated, determines third transformation matrix, and using the third transformation matrix as the mesh
Mark transformation matrix.
Identical with the method for above-mentioned the first transformation matrix of determination and the second transformation matrix in this step, computer equipment can
With preset based on anatomic landmarks point with reference to anatomic landmarks point set and the input of the second floating anatomic landmarks point set by second
Registration model, to obtain third transformation matrix, after obtaining third transformation matrix, computer equipment can directly by this
Three transformation matrixs are as object transformation matrix.
S30422 carries out image registration to the reference picture and the floating image according to the object transformation matrix.
Specifically, matrix and mesh that computer equipment can be constituted according to the coordinate position of each pixel of floating image
The methods of mark the product of transformation matrix, and combine interpolation method such as neighbour's interpolation, bilinear interpolation or Tri linear interpolation, label is floating
Motion video is mapped under label reference picture space, to realize label reference picture and label floating image under anatomical structure
Alignment, thus complete to label reference picture and mark floating image image registration.
Optionally, above-mentioned preset ratio and preset distance threshold can be adjusted according to such as under type: to be registered
Each mark point plus noise in reference picture and floating image, to ginseng to be registered in the way of the registration in above three stage
It examines image and floating image is registrated, obtain new object transformation matrix, new object transformation matrix is recycled, to above-mentioned ginseng
It examines image and floating image carries out image registration, and according to obtained registration result, utilize preset similarity measurement model, meter
The similarity measure values after being registrated between reference picture and floating image are calculated, according to the similarity measure values and preset similitude
Metric threshold is compared, if it is less than preset similarity measurement threshold value, then adjust above-mentioned preset ratio and it is preset away from
From at least one of threshold value, until finally obtained similarity measure values are greater than preset similarity measurement threshold value, from
And preset ratio and preset distance threshold are adjusted to suitably to be worth, and then the preset ratio using adjustment can be made
The registration accuracy for the image being registrated with the algorithm model of preset threshold value is higher.It should be noted that above-mentioned addition is made an uproar
Mean value, variance and the number of sound can be randomly provided.
Method for registering images provided in this embodiment, the reference subject to registration of the available label reference picture of computer equipment
The floating anatomic landmarks point set subject to registration of anatomic landmarks point set and label floating image;And anatomy is referred to according to subject to registration
Point set, floating anatomic landmarks point set subject to registration and the registration model based on anatomic landmarks point are marked, divides three phases to mark
Remember that reference picture and label floating image carry out image registration, each stage, which utilizes, passes through the such as preset ratio of certain condition
Mark point in interior mark point or preset distance threshold carries out image registration, rather than carries out image with whole mark points and match
Standard substantially reduces calculation amount, improves with Quasi velosity;In addition, the label point set in each stage is different, so as to reduce
Part anatomic landmarks point may be influenced the influence of accuracy of registration by erroneous detection, and the mark point in each stage is basis
Preset ratio or preset distance threshold etc. carry out the mark point that can be improved registration accuracy that screening is determined, therefore, this
The precision of image registration can be improved in the mode being registrated stage by stage that embodiment provides.
When above-mentioned target image registration model is the image registration model based on segmentation, computer equipment can use figure
The method for registering images that another embodiment shown in 4 provides carries out image to above-mentioned label reference picture and label floating image
Registration.Cut zone and corresponding image registration mould based on segmentation of the present embodiment what is involved is computer equipment according to extraction
Type carries out the realization process of image registration to above-mentioned label reference picture and label floating image.On the basis of above-described embodiment
On, optionally, the optional implementation of the another kind of above-mentioned S208 may include:
S402, obtains the corresponding segmentation reference picture of label reference picture and the corresponding segmentation of the floating image is floating
Motion video.
Specifically, above-mentioned segmentation reference picture and segmentation floating image can according to above-mentioned preset trained mind
The corresponding image after network model carries out Semantic features extraction to above-mentioned reference picture to be registered and floating image.It is optional
, computer equipment can use above-mentioned preset trained neural network model to reference picture to be registered and floating
Image carries out the segmentation of arbitrary region, to obtain segmentation reference picture and divide floating image.
S404, according to the segmentation reference picture, the segmentation floating image and the image registration mould based on segmentation
Type carries out image registration to the reference picture and the floating image.
Specifically, the above-mentioned image registration model based on segmentation can be equal for surface matching algorithm, mutual information method and gray scale
Any one in the corresponding algorithm model of the method for registering such as variance method.Computer equipment can be according to the segmentation of acquisition with reference to figure
Picture, segmentation floating image and the above-mentioned image registration model based on segmentation, determine Target Segmentation transformation matrix, thus according to this
Above-mentioned floating image to be registered is mapped under the space coordinate of reference picture by Target Segmentation transformation matrix, is completed with reference to figure
The registration of picture and floating image.
Method for registering images provided in this embodiment, the corresponding segmentation ginseng of the available label reference picture of computer equipment
Examine image and the corresponding segmentation floating image of floating image;And according to segmentation reference picture, segmentation floating image and based on segmentation
Image registration model, image registration is carried out to reference picture and floating image.In the present embodiment, computer equipment can basis
The segmentation reference picture obtained after Semantic features extraction and segmentation floating image are carried out, the preset figure based on segmentation is directly utilized
Picture registration model carries out image registration to reference picture and floating image, and implementation is simpler.
Fig. 5 is the method for registering images that another embodiment provides.What is involved is computer equipments according to upper for the present embodiment
The registration result obtained after embodiment is registrated reference picture and floating image is stated, using preset image integration model,
The process of image integration is carried out to the registration result.On the basis of the above embodiments, optionally, the above method can also wrap
It includes:
S502 is obtained and is carried out the registration result after image registration to the reference picture and the floating image.
In this step, above-mentioned registration result is to match to what is obtained after above-mentioned reference picture and floating image progress image registration
Reference picture and floating image after standard.
S504 carries out image integration to the registration result according to the registration result and preset image integration model.
In this step, above-mentioned preset image integration model can be in the methods of Tri linear interpolation and B-spline interpolation
Any one.Image integration can be the registration figure that more than two width or two width will be obtained from different imaging devices or different moments
Picture organically combines each image using certain algorithm.Computer equipment can use preset image integration mould
Type, by above-mentioned registration result reference picture and floating image integrate, to obtain floating image under reference picture space
The warp image combined with reference picture.
Method for registering images provided in this embodiment, computer equipment is available to carry out reference picture and floating image
Registration result after image registration;To carry out image to registration result according to registration result and preset image integration model
Reference picture and floating image are integrated into piece image by integration with realizing, thus by ground complementary the advantages of each image
Organically combine, to obtain the richer new images of information content, so that preferably auxiliary doctor utilizes the image after integration
The case where judging patient.
Fig. 6 is the method for registering images flow diagram that another embodiment provides.What is involved is computers for the present embodiment
The objective matrix that equipment is obtained according to above-described embodiment, and the image after down-sampling is carried out to reference picture and floating image,
Using gradient descent method, similarity measure values are adjusted, to determine the realization process of target component.On the basis of above-described embodiment
On, optionally, the above method can also include:
S602 obtains the object transformation matrix.
S604 according to the object transformation matrix, carries out the down-sampling obtained after down-sampling operation to the reference picture
Reference picture and obtained down-sampling floating image after down-sampling operation is carried out to the floating image, determines the down-sampling ginseng
Examine the similarity measure values between image and the corresponding transformed floating image of the down-sampling floating image.
Specifically, after computer equipment can obtain down-sampling to above-mentioned reference picture and floating image progress down-sampling
Down-sampling reference picture and down-sampling floating image, optionally, under being carried out once to above-mentioned reference picture and floating image
Sampling operation obtains down-sampling reference picture and down-sampling floating image, and floating to down-sampling using above-mentioned object transformation matrix
Motion video carries out spatial alternation, obtains transformed floating image, and then utilize the computation model of preset similarity measure values
Such as the corresponding algorithm model of the methods of mutual information method, gray scale average variance method, determine that the transformed floating image and down-sampling are joined
Examine the similarity measure values between image.
S606 carries out in translation, rotation process, tilt operation and zoom operations extremely the object transformation matrix
A few operation, extracts the corresponding initial parameter of the object transformation matrix.
Specifically, corresponding object transformation matrix can be 4*4 if reference picture and floating image are 3-D image
Matrix, computer equipment can to above-mentioned object transformation matrix carry out translation, rotation process, tilt operation and scaling grasp
Make, by object transformation matrix decomposition be translation matrix, spin matrix, tilt four 4*4 such as matrix and scaled matrix matrix, into
And the translation distance according to the matrix of four 4*4 under three-dimensional system of coordinate, rotation angle, tilt angle and scaling respectively
Deng obtaining the corresponding initial parameter of 12 object transformation matrixes.Similar, if reference picture and floating image are two dimensional image,
The then corresponding initial parameter of the available 8 object transformation matrixes of computer equipment.
S608 determines that target is joined according to the similarity measure values, the initial parameter and preset gradient descent method
Number.
Specifically, computer equipment can adjust above-mentioned initial parameter according to preset gradient descent method, so that above-mentioned
Similarity measure values are optimal, and using the corresponding parameter adjusted of optimal similarity measure values as target component.It can
Choosing, computer equipment can determine the corresponding final transformation matrix of the target component according to target component, and final using this
Transformation matrix is registrated reference picture and floating image.
Optionally, computer equipment can also carry out multiple down-sampling operation to above-mentioned reference picture and floating image, than
As carried out down-sampling three times and respectively obtaining corresponding down-sampling reference picture and down-sampling floating image.Further, it is adopted under
Sample reference picture may include the corresponding first down-sampling reference picture of first time down-sampling, second of down-sampling corresponding second
Down-sampling reference picture and the corresponding third down-sampling reference picture of third time down-sampling, similar, down-sampling floating image can
To include the corresponding first down-sampling floating image of first time down-sampling, the corresponding second down-sampling floating figure of second of down-sampling
Picture and the corresponding third down-sampling floating image of third time down-sampling.At this point it is possible to the sharp target component that determines with the following method: the
One step: computer equipment can use object transformation matrix and carry out spatial alternation to third down-sampling floating image, make its mapping
To under the corresponding space coordinates of third down-sampling reference picture, transformed third floating image is obtained, and utilize preset
The computation model of similarity measure values determines transformed third floating image between third down-sampling reference picture
One similarity measure values;Step 2: computer equipment can use preset gradient descent method adjust above-mentioned initial parameter so that
It obtains first similarity metric to be optimal, and new target is determined according to the optimal corresponding parameter of first similarity metric
Transformation matrix, and the second down-sampling floating image and down-sampling reference picture are continued to execute to using new object transformation square
The operation of the first step and second step is stated, until having executed the above-mentioned first step and second step to initial reference picture and floating image
Operation, using the corresponding parameter of finally obtained optimal similarity measure values as target component, so that computer equipment
The corresponding final transformation matrix of the target component can be determined according to target component, and using the final transformation matrix to reference to figure
Picture and floating image are registrated.
Optionally, computer equipment can be first with the corresponding image integration method of embodiment shown in fig. 5, to above-mentioned ginseng
It examines image and above-mentioned floating image carries out the registration result after image registration and carries out image integration, recycle provided in this embodiment
Embodiment illustrated in fig. 5 is obtained using the registration result that final transformation matrix is registrated reference picture and floating image
The result of integration optimize, also can use image optimization method provided in this embodiment to above-mentioned reference picture and above-mentioned
Floating image carries out the registration result after image registration and carries out image optimization, recycles the corresponding image of embodiment shown in fig. 5
Integration method carries out the present embodiment using the registration result that final transformation matrix is registrated reference picture and floating image
Image integration, the present embodiment is to this and without limitation.
Method for registering images provided in this embodiment, the available object transformation matrix of computer equipment, and according to target
Transformation matrix carries out the down-sampling reference picture obtained after down-sampling operation to reference picture and carries out down-sampling to floating image
The down-sampling floating image obtained after operation determines that down-sampling reference picture and down-sampling floating image are corresponding transformed floating
Similarity measure values between motion video;Translation, rotation process, tilt operation and scaling behaviour are carried out to object transformation matrix
The corresponding initial parameter of object transformation matrix is extracted in the operation of at least one of work;And then according to similarity measure values, initial ginseng
Several and preset gradient descent method determines target component, since target component is the corresponding parameter of optimal similarity measure values,
It therefore, is also preferably, to utilize the final transformation matrix in this way according to the final transformation matrix that the target component is determined, to floating
The precision that motion video and reference picture are registrated is also higher, further improves the precision of image registration.
It is following by a simply example, to introduce the process of the embodiment of the present application method for registering images.It specifically can be with
Referring to following steps:
S702, computer equipment obtain reference picture and floating image to be registered.
S704, computer equipment carry out the extraction of semantic information to the reference picture and the floating image, are wrapped
Include the label reference picture and label floating image of institute's semantic information;Institute's semantic information includes: point of the floating image
Cut cutting in region and anatomy mark point extremely at least one of region and anatomy mark point and the reference picture
It is one few.
S706, computer equipment determine the label ginseng according to institute's semantic information from preset image registration model
Examine image and the corresponding target image registration model of the label floating image;The preset image registration model includes
Image registration model based on segmentation and the registration model based on anatomic landmarks point.
S708, computer equipment judge whether above-mentioned target image registration model is the matching based on anatomic landmarks point
Quasi-mode type, if so, S710 is continued to execute, if it is not, executing S740.
S710, computer equipment obtain the subject to registration of label reference picture and refer to anatomic landmarks point set and the mark
Remember the floating anatomic landmarks point set subject to registration of floating image.
S712, computer equipment is according to described subject to registration with reference to anatomic landmarks point set and the floating anatomy subject to registration
The matching result for marking point set determines described subject to registration with reference to anatomic landmarks point set and the floating anatomic landmarks subject to registration
The identical mark point intersection of the title of mark point in point set, and choose described subject to registration with reference to described in anatomic landmarks point concentration
Mark point intersection is as initial reference anatomic landmarks point set, and chooses the floating anatomic landmarks point subject to registration and concentrate institute
Mark point intersection is stated as initial floating anatomic landmarks point set.
S714, computer equipment is according to the initial reference anatomic landmarks point set, the initial floating anatomic landmarks
Point set and the registration model based on anatomic landmarks point, determine first registration result.
S716, computer equipment determine in the preset ratio according to the first space length set and preset ratio
The corresponding first floating anatomic landmarks point set of the first space length;Wherein, recording in the first space length set has
First space subject to registration that each correspondence markings point is concentrated with reference to anatomic landmarks point set and the first registration result point
Distance.
S718, computer equipment judge whether the number for the mark point that the first floating anatomic landmarks point is concentrated is less than
The preset quantity threshold, if so, S720 is continued to execute, if it is not, then executing S722.
S720, computer equipment is using first transformation matrix as object transformation matrix.
S722, computer equipment obtain described subject to registration with reference to anatomic landmarks point concentration and the first floating anatomy
Point set corresponding first is marked to refer to anatomic landmarks point set.
S724, computer equipment refer to anatomic landmarks point set, the first floating anatomic landmarks according to described first
Point set and the registration model based on anatomic landmarks point, determine the second transformation matrix.
S726, computer equipment is according to second transformation matrix and the floating anatomic landmarks point set subject to registration, really
Fixed second registration result point set.
S728, computer equipment determine according to second space distance set and preset distance threshold and are less than described preset
Distance threshold second space apart from corresponding second floating anatomic landmarks point set;Remember in the second space distance set
Record, which has, described subject to registration concentrates the second of each correspondence markings point with reference to anatomic landmarks point set and the second registration result point
Space length.
S730, computer equipment judge whether the number for the mark point that the second floating anatomic landmarks point is concentrated is less than
The preset threshold number, if so, S732 is continued to execute, if it is not, then executing S734.
S732, computer equipment is using second transformation matrix as the object transformation matrix.
S734, computer equipment obtain described subject to registration with reference to anatomic landmarks point concentration and the second floating anatomy
Point set corresponding second is marked to refer to anatomic landmarks point set.
S736, computer equipment refer to anatomic landmarks point set, the second floating anatomic landmarks according to described second
Point set and the registration model based on anatomic landmarks point determine third transformation matrix, and the third transformation matrix are made
For the object transformation matrix.
S738, computer equipment carry out the reference picture and the floating image according to the object transformation matrix
Image registration;After having executed S738, S744 is continued to execute.
S740, computer equipment obtain the corresponding segmentation reference picture of label reference picture and the floating image pair
The segmentation floating image answered.
S742, computer equipment is according to the segmentation reference picture, the segmentation floating image and described based on segmentation
Image registration model carries out image registration to the reference picture and the floating image.
S744, computer equipment, which is obtained, carries out the registration knot after image registration to the reference picture and the floating image
Fruit.
S746, computer equipment according to the registration result and preset image integration model, to the registration result into
Row image integration.
S748, computer equipment obtain the object transformation matrix.
S750, computer equipment according to the object transformation matrix, to the reference picture carry out down-sampling operation after
To down-sampling reference picture and obtained down-sampling floating image after down-sampling operation is carried out to the floating image, determine institute
State the similarity measure values between down-sampling reference picture and the corresponding transformed floating image of the down-sampling floating image.
S752, computer equipment carry out translation, rotation process, tilt operation and scaling to the object transformation matrix
At least one of operation operation, extracts the corresponding initial parameter of the object transformation matrix.
S754, computer equipment is according to the similarity measure values, the initial parameter and preset gradient descent method, really
Set the goal parameter.
The working principle and technical effect of method for registering images provided in this embodiment are as described in above-described embodiment, herein not
It repeats again.
It should be understood that although each step in the flow chart of Fig. 2 to Fig. 6 is successively shown according to the instruction of arrow,
But these steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly state otherwise herein, these
There is no stringent sequences to limit for the execution of step, these steps can execute in other order.Moreover, Fig. 2 is into Fig. 6
At least part step may include that perhaps these sub-steps of multiple stages or stage are not necessarily same to multiple sub-steps
One moment executed completion, but can execute at different times, and the execution in these sub-steps or stage sequence is also not necessarily
Be successively carry out, but can at least part of the sub-step or stage of other steps or other steps in turn or
Alternately execute.
Fig. 7 is the image registration device structural schematic diagram that one embodiment provides.As shown in fig. 7, the apparatus may include
First obtains module 702, the first extraction module 704, the first determining module 706 and registration module 708.
Specifically, first obtains module 702, for obtaining reference picture and floating image to be registered;
First extraction module 704 is obtained for carrying out the extraction of semantic information to the reference picture and the floating image
To the label reference picture and label floating image for including institute's semantic information;
First determining module 706, for determining the mark from preset image registration model according to institute's semantic information
Remember reference picture and the corresponding target image registration model of the label floating image;
Registration module 708 is used for according to institute's semantic information and the target image registration model, to the reference picture
Image registration is carried out with the floating image.
Optionally, institute's semantic information include: in the cut zone and anatomy mark point of the floating image at least
One and the reference picture cut at least one of region and anatomy mark point;The preset image registration mould
Type includes the image registration model based on segmentation and the registration model based on anatomic landmarks point.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
In the image registration device that another embodiment provides, on the basis of above-mentioned embodiment illustrated in fig. 7, when described
When target image registration model is the registration model based on anatomic landmarks point, optionally, above-mentioned registration module 708 can be with
Including first acquisition unit and the first registration unit.
Specifically, first acquisition unit, refers to anatomic landmarks point for obtaining the subject to registration of label reference picture
The floating anatomic landmarks point set subject to registration of collection and the label floating image;
First registration unit, for according to described subject to registration with reference to anatomic landmarks point set, floatings subject to registration dissection
Label point set and the registration model based on anatomic landmarks point are learned, figure is carried out to the reference picture and the floating image
As registration.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
In the image registration device that another embodiment provides, on the basis of the above embodiments, optionally, above-mentioned the
One registration unit may include the first determining subelement, the second determining subelement and registration subelement.
Specifically, first determine subelement, for root according to it is described it is subject to registration with reference to anatomic landmarks point set and it is described to
Registration floating anatomic landmarks point concentrates the matching result of the title of each mark point, determines mark point intersection;
Second determination subelement is used for according to the mark point intersection, from the reference anatomic landmarks point set subject to registration
It is concentrated with the floating anatomic landmarks point subject to registration and determines initial reference anatomic landmarks point set and initial dissection of floating respectively
Learn label point set;
It is registrated subelement, for according to the initial reference anatomic landmarks point set, the initial floating anatomic landmarks
Point set and the registration model based on anatomic landmarks point carry out image to the reference picture and the floating image and match
It is quasi-.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
In the image registration device structure that another embodiment provides, on the basis of the above embodiments, optionally, on
Stating registration module 708 can also include second acquisition unit and the second registration unit.
Second acquisition unit, for obtaining the corresponding segmentation reference picture of label reference picture and the floating image
Corresponding segmentation floating image;
Second registration unit, for according to the segmentation reference picture, the segmentation floating image and it is described be based on divide
Image registration model, image registration is carried out to the reference picture and the floating image.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
Fig. 8 is the image registration device structural schematic diagram that another embodiment provides.On the basis of the above embodiments, may be used
Choosing, above-mentioned apparatus can also include the second acquisition module 710 and integrate module 712.
Second obtains module 710, for obtaining to after the reference picture and floating image progress image registration
Registration result;
Integrate module 712, for according to the registration result and preset image integration model, to the registration result into
Row image integration.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
Fig. 9 is the image registration device structural schematic diagram that another embodiment provides.On the basis of the above embodiments, may be used
Choosing, above-mentioned apparatus can also include that third obtains module 714, the second determining module 716, the second extraction module 718 and third
Determining module 720.
Third obtains module 714, for obtaining the object transformation matrix.
Second determining module 716, for carrying out down-sampling operation according to the object transformation matrix, to the reference picture
The down-sampling reference picture that obtains afterwards and the down-sampling floating image obtained after down-sampling operation is carried out to the floating image, really
Similarity measurements between the fixed down-sampling reference picture and the corresponding transformed floating image of the down-sampling floating image
Magnitude;
Second extraction module 718, for carrying out translation, rotation process, tilt operation to the object transformation matrix
With the operation of at least one of zoom operations, the corresponding initial parameter of the object transformation matrix is extracted;
Third determining module 720, for being declined according to the similarity measure values, the initial parameter and preset gradient
Method determines target component.
Image registration device provided in this embodiment can execute above method embodiment, realization principle and technology effect
Seemingly, details are not described herein for fruit.
In one embodiment, a kind of computer equipment, including memory and processor are provided, is stored in memory
Computer program, the processor perform the steps of when executing computer program
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtains including institute's semantic information
Mark reference picture and label floating image;
According to institute's semantic information, the label reference picture and the label are determined from preset image registration model
The corresponding target image registration model of floating image;
According to institute's semantic information and the target image registration model, to the reference picture and the floating image into
Row image registration.
Computer equipment provided by the above embodiment, implementing principle and technical effect are similar with above method embodiment,
Details are not described herein.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtains including institute's semantic information
Mark reference picture and label floating image;
According to institute's semantic information, the label reference picture and the label are determined from preset image registration model
The corresponding target image registration model of floating image;
According to institute's semantic information and the target image registration model, to the reference picture and the floating image into
Row image registration.
Computer readable storage medium provided by the above embodiment, implementing principle and technical effect and the above method are implemented
Example is similar, and details are not described herein.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of embodiment described above can be combined arbitrarily, for simplicity of description, not to above-mentioned reality
It applies all possible combination of each technical characteristic in example to be all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of method for registering images, which is characterized in that the described method includes:
Obtain reference picture and floating image to be registered;
The extraction that semantic information is carried out to the reference picture and the floating image, obtain include institute's semantic information label
Reference picture and label floating image;
According to institute's semantic information, determine that the label reference picture and the label float from preset image registration model
The corresponding target image registration model of image;
According to institute's semantic information and the target image registration model, figure is carried out to the reference picture and the floating image
As registration.
2. the method according to claim 1, wherein institute's semantic information includes: the segmentation of the floating image
In the cut zone and anatomy mark point of at least one of region and anatomy mark point and the reference picture extremely
It is one few;The preset image registration model includes the image registration model based on segmentation and matching based on anatomic landmarks point
Quasi-mode type.
3. according to the method described in claim 2, it is characterized in that, when the target image registration model is described based on dissection
It is described according to institute's semantic information and the target image registration model when learning the registration model of mark point, to described with reference to figure
Picture and the floating image carry out image registration, comprising:
Obtain it is described label reference picture it is subject to registration with reference to anatomic landmarks point set and it is described label floating image it is subject to registration
Floating anatomic landmarks point set;
According to it is described it is subject to registration with reference to anatomic landmarks point set, the floating anatomic landmarks point set subject to registration and it is described based on solution
The registration model for learning mark point is cutd open, image registration is carried out to the reference picture and the floating image.
4. according to the method described in claim 3, it is characterized in that, described according to described subject to registration with reference to anatomic landmarks point
Collection, the floating anatomic landmarks point set subject to registration and the registration model based on anatomic landmarks point, to described with reference to figure
Picture and the floating image carry out image registration, comprising:
Each label is concentrated according to the reference anatomic landmarks point set subject to registration and the floating anatomic landmarks point subject to registration
The matching result of the title of point, determines mark point intersection;
According to the mark point intersection, from reference anatomic landmarks point set and the floating anatomy mark subject to registration subject to registration
Note point is concentrated determines initial reference anatomic landmarks point set and initial floating anatomic landmarks point set respectively;
According to the initial reference anatomic landmarks point set, the initial floating anatomic landmarks point set and it is described be based on anatomy
The registration model of mark point carries out image registration to the reference picture and the floating image.
5. according to the method described in claim 2, it is characterized in that, when the target image registration model is described based on segmentation
Image registration model when, then it is described according to institute's semantic information and the target image registration model, to the reference picture
Image registration is carried out with the floating image, comprising:
Obtain the corresponding segmentation reference picture of label reference picture and the corresponding segmentation floating image of the floating image;
According to the segmentation reference picture, the segmentation floating image and the image registration model based on segmentation, to described
Reference picture and the floating image carry out image registration.
6. according to the method described in claim 2, it is characterized in that, the method also includes:
It obtains and carries out the registration result after image registration to the reference picture and the floating image;
According to the registration result and preset image integration model, image integration is carried out to the registration result.
7. according to the method described in claim 6, it is characterized in that, it is described to the reference picture and the floating image into
After row image registration, the method also includes:
Obtain the object transformation matrix;
According to the object transformation matrix, the reference picture is carried out after down-sampling operation obtained down-sampling reference picture and
The down-sampling floating image obtained after down-sampling operation is carried out to the floating image, determines the down-sampling reference picture and institute
State the similarity measure values between the corresponding transformed floating image of down-sampling floating image;
At least one of translation, rotation process, tilt operation and zoom operations behaviour are carried out to the object transformation matrix
Make, extracts the corresponding initial parameter of the object transformation matrix;
According to the similarity measure values, the initial parameter and preset gradient descent method, target component is determined.
8. a kind of image registration device, which is characterized in that described device includes:
First obtains module, for obtaining reference picture and floating image to be registered;
First extraction module, for the extraction to the reference picture and floating image progress semantic information, including
The label reference picture and label floating image of institute's semantic information;
First determining module, for according to institute's semantic information, determining the label reference from preset image registration model
Image and the corresponding target image registration model of the label floating image;
Registration module, for according to institute's semantic information and the target image registration model, to the reference picture and described
Floating image carries out image registration.
9. a kind of computer equipment, the computer equipment includes memory and processor, and the memory is stored with computer
Program, which is characterized in that the processor realizes any one of claim 1-7 the method when executing the computer program
The step of.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811637721.8A CN109754396B (en) | 2018-12-29 | 2018-12-29 | Image registration method and device, computer equipment and storage medium |
CN202110189069.3A CN112967236B (en) | 2018-12-29 | 2018-12-29 | Image registration method, device, computer equipment and storage medium |
PCT/CN2019/127695 WO2020135374A1 (en) | 2018-12-25 | 2019-12-24 | Image registration method and apparatus, computer device and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811637721.8A CN109754396B (en) | 2018-12-29 | 2018-12-29 | Image registration method and device, computer equipment and storage medium |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110189069.3A Division CN112967236B (en) | 2018-12-29 | 2018-12-29 | Image registration method, device, computer equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109754396A true CN109754396A (en) | 2019-05-14 |
CN109754396B CN109754396B (en) | 2021-02-19 |
Family
ID=66404444
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110189069.3A Active CN112967236B (en) | 2018-12-29 | 2018-12-29 | Image registration method, device, computer equipment and storage medium |
CN201811637721.8A Active CN109754396B (en) | 2018-12-25 | 2018-12-29 | Image registration method and device, computer equipment and storage medium |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110189069.3A Active CN112967236B (en) | 2018-12-29 | 2018-12-29 | Image registration method, device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (2) | CN112967236B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110415279A (en) * | 2019-06-25 | 2019-11-05 | 北京全域医疗技术集团有限公司 | Method for registering images, device and equipment |
CN110473233A (en) * | 2019-07-26 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Method for registering, computer equipment and storage medium |
CN110766730A (en) * | 2019-10-18 | 2020-02-07 | 上海联影智能医疗科技有限公司 | Image registration and follow-up evaluation method, storage medium and computer equipment |
WO2020135374A1 (en) * | 2018-12-25 | 2020-07-02 | 上海联影智能医疗科技有限公司 | Image registration method and apparatus, computer device and readable storage medium |
CN112750152A (en) * | 2021-01-18 | 2021-05-04 | 上海联影医疗科技股份有限公司 | Image registration method and device, computer equipment and storage medium |
CN112766314A (en) * | 2020-12-31 | 2021-05-07 | 上海联影智能医疗科技有限公司 | Anatomical structure recognition method, electronic device, and storage medium |
CN113616350A (en) * | 2021-07-16 | 2021-11-09 | 元化智能科技(深圳)有限公司 | Verification method and device for selected positions of marking points, terminal equipment and storage medium |
CN114187335A (en) * | 2021-11-09 | 2022-03-15 | 北京东软医疗设备有限公司 | Registration method, device and equipment for multi-view medical image |
WO2022051977A1 (en) * | 2020-09-10 | 2022-03-17 | 西安大医集团股份有限公司 | Image registration method and device |
CN117115220A (en) * | 2023-08-31 | 2023-11-24 | 阿里巴巴达摩院(杭州)科技有限公司 | Image processing method, service providing method, device, equipment and storage medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024067829A1 (en) * | 2022-09-30 | 2024-04-04 | Shanghai United Imaging Healthcare Co., Ltd. | Methods, systems, and storage medium for image registration |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101126722A (en) * | 2007-09-30 | 2008-02-20 | 西北工业大学 | Cone-beam CT beam hardening calibration method based on registration model emulation |
CN101241601A (en) * | 2008-02-19 | 2008-08-13 | 深圳先进技术研究院 | Graphic processing joint center parameter estimation method |
CN103337065A (en) * | 2013-05-22 | 2013-10-02 | 西安电子科技大学 | Non-rigid registering method of mouse three-dimensional CT image |
CN103793915A (en) * | 2014-02-18 | 2014-05-14 | 上海交通大学 | Low-cost mark-free registration system and method in neurosurgery navigation |
CN104160424A (en) * | 2012-03-08 | 2014-11-19 | 皇家飞利浦有限公司 | Intelligent landmark selection to improve registration accuracy in multimodal image fusion |
CN104392426A (en) * | 2014-10-23 | 2015-03-04 | 华中科技大学 | Adaptive markerless three-dimensional point cloud automatic registration method |
CN106920234A (en) * | 2017-02-27 | 2017-07-04 | 北京连心医疗科技有限公司 | A kind of method of the automatic radiotherapy planning of combined type |
US20180144463A1 (en) * | 2016-11-21 | 2018-05-24 | International Business Machines Corporation | Slice selection for interpolation-based 3d manual segmentation |
Family Cites Families (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101471646B1 (en) * | 2012-08-21 | 2014-12-26 | 인하대학교 산학협력단 | Parallel processing of 3D medical image registration by GP-GPU |
CN103345741B (en) * | 2013-06-13 | 2016-07-06 | 华中科技大学 | A kind of non-rigid multi modal medical image Precision Registration |
CN104287830A (en) * | 2013-07-18 | 2015-01-21 | 中国科学院深圳先进技术研究院 | Intraoperative real-time registration method based on Kinect camera |
CN105303547A (en) * | 2014-07-11 | 2016-02-03 | 东北大学 | Multiphase CT image registration method based on grid matching Demons algorithm |
CN104867104B (en) * | 2015-05-20 | 2017-12-15 | 天津大学 | Target mouse anatomical structure collection of illustrative plates acquisition methods based on the non-rigidity registration of XCT images |
CN106611411B (en) * | 2015-10-19 | 2020-06-26 | 上海联影医疗科技有限公司 | Method for segmenting ribs in medical image and medical image processing device |
US20170337682A1 (en) * | 2016-05-18 | 2017-11-23 | Siemens Healthcare Gmbh | Method and System for Image Registration Using an Intelligent Artificial Agent |
CN106934821B (en) * | 2017-03-13 | 2020-06-23 | 中国科学院合肥物质科学研究院 | Conical beam CT and CT image registration method based on ICP algorithm and B spline |
CN107049475A (en) * | 2017-04-19 | 2017-08-18 | 纪建松 | Liver cancer local ablation method and system |
CN107341824B (en) * | 2017-06-12 | 2020-07-28 | 西安电子科技大学 | Comprehensive evaluation index generation method for image registration |
CN107451983A (en) * | 2017-07-18 | 2017-12-08 | 中山大学附属第六医院 | The three-dimensional fusion method and system of CT images |
CN107610162A (en) * | 2017-08-04 | 2018-01-19 | 浙江工业大学 | A kind of three-dimensional multimode state medical image autoegistration method based on mutual information and image segmentation |
CN107578434A (en) * | 2017-08-25 | 2018-01-12 | 上海嘉奥信息科技发展有限公司 | VR rendering intents and system based on 3D point cloud rapid registering |
-
2018
- 2018-12-29 CN CN202110189069.3A patent/CN112967236B/en active Active
- 2018-12-29 CN CN201811637721.8A patent/CN109754396B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101126722A (en) * | 2007-09-30 | 2008-02-20 | 西北工业大学 | Cone-beam CT beam hardening calibration method based on registration model emulation |
CN101241601A (en) * | 2008-02-19 | 2008-08-13 | 深圳先进技术研究院 | Graphic processing joint center parameter estimation method |
CN104160424A (en) * | 2012-03-08 | 2014-11-19 | 皇家飞利浦有限公司 | Intelligent landmark selection to improve registration accuracy in multimodal image fusion |
CN103337065A (en) * | 2013-05-22 | 2013-10-02 | 西安电子科技大学 | Non-rigid registering method of mouse three-dimensional CT image |
CN103793915A (en) * | 2014-02-18 | 2014-05-14 | 上海交通大学 | Low-cost mark-free registration system and method in neurosurgery navigation |
CN104392426A (en) * | 2014-10-23 | 2015-03-04 | 华中科技大学 | Adaptive markerless three-dimensional point cloud automatic registration method |
US20180144463A1 (en) * | 2016-11-21 | 2018-05-24 | International Business Machines Corporation | Slice selection for interpolation-based 3d manual segmentation |
CN106920234A (en) * | 2017-02-27 | 2017-07-04 | 北京连心医疗科技有限公司 | A kind of method of the automatic radiotherapy planning of combined type |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020135374A1 (en) * | 2018-12-25 | 2020-07-02 | 上海联影智能医疗科技有限公司 | Image registration method and apparatus, computer device and readable storage medium |
CN110415279A (en) * | 2019-06-25 | 2019-11-05 | 北京全域医疗技术集团有限公司 | Method for registering images, device and equipment |
CN110473233B (en) * | 2019-07-26 | 2022-03-01 | 上海联影智能医疗科技有限公司 | Registration method, computer device, and storage medium |
CN110473233A (en) * | 2019-07-26 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Method for registering, computer equipment and storage medium |
CN110766730A (en) * | 2019-10-18 | 2020-02-07 | 上海联影智能医疗科技有限公司 | Image registration and follow-up evaluation method, storage medium and computer equipment |
CN110766730B (en) * | 2019-10-18 | 2023-02-28 | 上海联影智能医疗科技有限公司 | Image registration and follow-up evaluation method, storage medium and computer equipment |
WO2022051977A1 (en) * | 2020-09-10 | 2022-03-17 | 西安大医集团股份有限公司 | Image registration method and device |
CN112766314A (en) * | 2020-12-31 | 2021-05-07 | 上海联影智能医疗科技有限公司 | Anatomical structure recognition method, electronic device, and storage medium |
CN112766314B (en) * | 2020-12-31 | 2024-05-28 | 上海联影智能医疗科技有限公司 | Anatomical structure recognition method, electronic device, and storage medium |
CN112750152A (en) * | 2021-01-18 | 2021-05-04 | 上海联影医疗科技股份有限公司 | Image registration method and device, computer equipment and storage medium |
CN113616350A (en) * | 2021-07-16 | 2021-11-09 | 元化智能科技(深圳)有限公司 | Verification method and device for selected positions of marking points, terminal equipment and storage medium |
CN114187335A (en) * | 2021-11-09 | 2022-03-15 | 北京东软医疗设备有限公司 | Registration method, device and equipment for multi-view medical image |
CN117115220A (en) * | 2023-08-31 | 2023-11-24 | 阿里巴巴达摩院(杭州)科技有限公司 | Image processing method, service providing method, device, equipment and storage medium |
CN117115220B (en) * | 2023-08-31 | 2024-04-26 | 阿里巴巴达摩院(杭州)科技有限公司 | Image processing method, service providing method, device, equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN112967236B (en) | 2024-02-27 |
CN109754396B (en) | 2021-02-19 |
CN112967236A (en) | 2021-06-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109754396A (en) | Method for registering, device, computer equipment and the storage medium of image | |
US4835688A (en) | Three-dimensional image processing apparatus | |
KR102450931B1 (en) | Image registration method and associated model training method, apparatus, apparatus | |
CN108369736A (en) | Method and system for the tissue volume for calculating excision according to image data in 2D/2.5D arts | |
CN107592802A (en) | Strengthen the system and method for guide abdominal videoendoscopic surgery process by anatomical model | |
CN110111305B (en) | Processing scheme generation method, device, equipment and storage medium | |
CN111080573A (en) | Rib image detection method, computer device and storage medium | |
CN113112559A (en) | Ultrasonic image segmentation method and device, terminal equipment and storage medium | |
CN104240180A (en) | Method and device for achieving automatic adjusting of images | |
US10078906B2 (en) | Device and method for image registration, and non-transitory recording medium | |
US20210256703A1 (en) | Machine learning device, estimation device, non-transitory computer readable medium, and learned model | |
CN108198235A (en) | A kind of three dimentional reconstruction method, apparatus, equipment and storage medium | |
US20190392552A1 (en) | Spine image registration method | |
CN111192268A (en) | Medical image segmentation model construction method and CBCT image bone segmentation method | |
CN117115068A (en) | Medical image processing method, medical image processing device, computer equipment and storage medium | |
CN110533120A (en) | Image classification method, device, terminal and the storage medium of organ tubercle | |
CN108597589B (en) | Model generation method, target detection method and medical imaging system | |
CN113129418B (en) | Target surface reconstruction method, device, equipment and medium based on three-dimensional image | |
CN112950684B (en) | Target feature extraction method, device, equipment and medium based on surface registration | |
CN111383236B (en) | Method, apparatus and computer-readable storage medium for labeling regions of interest | |
Li et al. | A vision-based navigation system with markerless image registration and position-sensing localization for oral and maxillofacial surgery | |
CN114093462A (en) | Medical image processing method, computer device, and storage medium | |
CN109685796B (en) | Medical image processing method, apparatus, device and storage medium | |
US11341661B2 (en) | Method and apparatus for registering live medical image with anatomical model | |
CN113962957A (en) | Medical image processing method, bone image processing method, device and equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20210812 Address after: Room 3674, 3 / F, 2879 Longteng Avenue, Xuhui District, Shanghai, 200232 Patentee after: SHANGHAI UNITED IMAGING INTELLIGENT MEDICAL TECHNOLOGY Co.,Ltd. Patentee after: Lianying intelligent medical technology (Beijing) Co.,Ltd. Address before: Room 3674, 3 / F, 2879 Longteng Avenue, Xuhui District, Shanghai, 200232 Patentee before: SHANGHAI UNITED IMAGING INTELLIGENT MEDICAL TECHNOLOGY Co.,Ltd. |